Camera Configuration Design in Cooperative Active Visual 3d Reconstruction: A Statistical Approach

摘要

Visual 3D reconstruction is an essential technique in computer vision which restores the 3D model of the scene from multi-view images. In this paper, we propose a statistical framework for the active visual 3D reconstruction. We first derive a closed-form expression to characterize the dependence of the reconstruction performance on 3D point’s position and camera setting parameters under the binocular camera setting. Then we derive several closed-form solutions and propose an efficient optimization algorithm to find the optimal camera configurations with the best reconstruction quality under various settings. Numerical results validate the performance of our methods in terms of both reconstruction accuracy and computational efficiency.